trainer_state.json

{
  "best_metric": 0.9655162692070007,
  "best_model_checkpoint": "./test-augment/checkpoint-20",
  "epoch": 2.0,
  "eval_steps": 500,
  "global_step": 20,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6375,
      "eval_loss": 1.1256365776062012,
      "eval_runtime": 1.0412,
      "eval_samples_per_second": 76.836,
      "eval_steps_per_second": 1.921,
      "step": 10
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7125,
      "eval_loss": 0.9655162692070007,
      "eval_runtime": 1.0636,
      "eval_samples_per_second": 75.217,
      "eval_steps_per_second": 1.88,
      "step": 20
    },
    {
      "epoch": 2.0,
      "step": 20,
      "total_flos": 5.091978035331072e+16,
      "train_loss": 1.3109167098999024,
      "train_runtime": 32.3348,
      "train_samples_per_second": 39.586,
      "train_steps_per_second": 0.619
    }
  ],
  "logging_steps": 100,
  "max_steps": 20,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 500,
  "total_flos": 5.091978035331072e+16,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}